Most children transition from presymbolic (e.g., gestures and vocalizations) to symbolic (e.g., words) communication during the first year of life. This transition enables children to communicate more efficiently and leads to further developments in language acquisition. The exact mechanisms that facilitate this transition are still not fully understood, although social-interactionist theories describe both endogenous and environmental factors affecting early word learning. The transition into symbolic communication is at risk for many children with developmental delays including autism. When delays persist into preschool ages, children are frequently taught to communicate with some form of augmentative or alternative communication (AAC). These forms may include signs, picture selection, and/or voice output communication aids (VOCA). Although use of AAC by preschool children is increasing, little evidence exists linking specific child and environmental variables to communication outcomes. Children learning AAC present with very different characteristics and histories but there is no existing evidence base to guide researchers or practitioners in differentially applying interventions or interpreting results based on these variables. We hypothesize that, of the many variables that could relate to communication outcomes, a child's representational capacity- as indicated by cognitive development, comprehension, symbolic play, and prelinguistic communication level- is a significant predictor of communication outcomes. Further, we hypothesize that the effects of this predictor are mediated by the amounts of supports and input offered by the social environment. Through a longitudinal study, we propose to test hypothesized relationships between predictor and mediating variables and three different outcomes- symbolic vocabulary development, communication success, and symbol substitutions during communication repairs. Participants will be 100 young children with developmental disabilities and 20 young children with autism spectrum disorders (ASD), all of whom will be learning to communicate with AAC. Separate analyses will be completed with the cohort of children with ASD to determine the extent to which our model applies to symbolic communication development by this important subgroup. The proposed research would produce the largest data set collected thus far from a prospective study of young children learning AAC. This research is likely to improve assessment and intervention practices for these children by identifying potential predictors of response to AAC interventions. Our focus on children learning AAC will inform theoretical accounts of symbolization by documenting the extent to which processes observed in learning speech also apply to learning other symbolic modes.